Differences

This shows you the differences between two versions of the page.

--- curriculum [2026/07/02 14:34] – [Period 1] respai-vic
+++ curriculum [2026/07/02 15:08] (current) – [Period 1] respai-vic
@@ Line 92: / Line 92: @@
 **CSC_53433_EP - Creative & Generative models in Computer Graphics (24h, 2 ECTS),  Marie-Paule Cani (EP), Julien Pettré (Inria)** (contact: marie-paule.cani@polytechnique.edu)
->  This course reviews content creation methods in modern computer graphics — ranging from static 3D shapes to animated landscapes and character motion  —  with a focus on their connection to AI.
+>  This course reviews the recent advances of content creation methods in computer graphics — ranging from the design of 3D shapes to animated landscapes and character motion  —  with a focus on their connection to AI. We first present expressive modeling methods (also called "smart 3D models"), knowledge-based or learned from examples, designed to help users create the 3D shapes they have in mind, and arrange them into complex static or animated scenes. We then discuss their recent evolution toward user-centric creative AI, leveraging deep generative models. Next, we examine the generation and control of character motion, from navigation tasks with the presentation of recent crowd models, to new trends for motor skills generation, such as the use of Deep Reinforcement Learning for training motion controllers of physically-based characters. These methods enable the creation of 3D agents capable of moving individually or in groups while interacting with their environment. Practical sessions are conducted using Unity and the C# language.
-We first introduce Expressive modeling methods, i.e. smart 3D models - either based on knowledge or trained from examples, designed to help users create/generate the 3D shapes their have in mind and organise them into complex, static or animated scenes. We then discuss their recent extension into user-centered Creative AI, relying on Deep generative models. Second, we study the use of AI in Character Animation, from motion planning methods to recent crowd navigation models and from the adaptation of kinematic gaits to deep reinforcement learning solutions used to generate the motor skills of physically-based characters. These methods result into 3D agents able to navigate alone or in groups while interacting with their environment. The lab sessions are held on Unity, based on C#.